tests of differences Flashcards
(21 cards)
Test of differences
An investigation of a hypothesis stating that two (or more) groups differ with respect to measures on a variable
Independent samples t-test
compares the means of two different groups to see if they are significantly different from each other
- must be two independent samples (e.g. different people in group A and B - comparing purchase intent between shoppers who saw add A and those who saw add B)
Paired samples t test
A technique used to test the hypothesis that mean scores differ on some interval or ratio scaled variable between related or paired samples
(compares mean of two related groups) e.g. measuring customer satisfaction before and after using new app feature
Test of means - does the analysis involve 2 groups or less?
t - test
Test of means - does the analysis involve 3 or more groups?
anova
Test of means - are you examining between group differences?
Independent samples
Test of means - are you examining within group differences?
related samples
variables measured - independent samples t test
one metric variable, one non metric variable with two groups
variables measured - paired samples t test
two metric variables from paired sample t test
variables measured - anova
one metric variable, one non metric variable with three or more groups
experimental group
exposed
control group
not exposed
pooled estimate of the standard error
an estimate of the standard error for a t test of independent means that assumes the variances of both groups are equal
Levene’s test for equality of variances
If significance (sig or p-value) value is greater than 0.05, then we can assume equal variances.
equally of means and significance value
if significance (sig or p-value) is less than 0.05 then the two means are significantly different
F test
Used to determine whether there is more variability in the scores of some sample than in the scores of another
- A larger ratio of variance between groups to varience within groups implies a greater value of F
= If F is large, the more likely it is that the differences in means has occurred as a result of the grouping variable
f ratio
if the calculated f ratio exceeds the critical f ratio, it implies that the results are statistically significant
- thus the null hypothesis (Ho) has to be rejected
Assumptions for parametric ANOVA test of differences
- The population (or sampling distribution) is normally distributed
- null = all group means are equal
null hypothesis of independent samples
the means of the two independent groups are equal
null hypothesis of paired samples
the mean difference between paired observations is 0
null hypothesis of anova
all group means are equal
